bases with symmetry signatures

Accepted Manuscript Title: A Theoretical Analysis of Ambivalent and Ambiphilic Lewis Acid/Bases with Symmetry Signatures Author: D. Michael P. Mingos PII: DOI: Reference:

S0010-8545(14)00329-4 http://dx.doi.org/doi:10.1016/j.ccr.2014.11.009 CCR 111962

To appear in:

Coordination Chemistry Reviews

Received date: Revised date: Accepted date:

31-8-2014 17-11-2014 27-11-2014

Please cite this article as: D.M.P. Mingos, A Theoretical Analysis of Ambivalent and Ambiphilic Lewis Acid/Bases with Symmetry Signatures, Coordination Chemistry Reviews (2014), http://dx.doi.org/10.1016/j.ccr.2014.11.009 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Manuscript

A Theoretical Analysis of Ambivalent and Ambiphilic Lewis Acid/Bases with Symmetry Signatures D. Michael P. Mingos* *

2014 ; accepted Available on line

cr

Received

ip t

Inorganic Chemistry laboratory, Oxford University, South Parks Road, Oxford OX1 3QR, UK, [email protected]

Ac

ce pt

ed

M

an

us

Abstract This review provides a theoretical underpinning of previously published definitions of ambidentate, ambivalent and ambiphilic ligands. The study encompasses ambivalent ligands such as NO, NR, N2R; ambiphilic molecules such as SO2, I2 and ambiphilic transition metal complexes, e.g. [Pt(PCy3)2]. These ambivalent molecules adopt alternative geometries which depend primarily on the number of electrons which they formally donate or accept. The theoretical analysis focuses initially on those complexes where the same ligand displays ambivalent properties within the same molecule in order to define the energetics of their interconversion. These square-pyramidal complexes provide a test-bed for generating data which throws light on the relative abilities of ambivalent ligands to adopt linear or bent geometries. The ligands were compared with NO and their relative abilities were placed in the following order PO>PH2>N2H>SO2>NO>NH2>NS. The linear nitrosyl ligand does not exert a trans- influence and this property has been contrasted with the nitrido- ligand which shows a large trans- influence. The conversion of NO to a non-linear geometry results in a strong trans- influence and this has significant catalytic and biological importance. Calculations on octahedral palladium complexes have been used to order the trans- influences of ambivalent ligands when they adopt their alternative symmetry signatures. The relative trans- influences are NO > PH2 > NS > N2H > NH2. The interconversion of linear and bent dinitrosyls provides an interesting inorganic example of valence tautomerism and this is noted as a general characteristic of ambivalent and ambiphilic ligands. The soft energy surface associated with these interconversions leads to the experimentally verified fluxional process. The energetics of adduct formation by ambiphilic ligands has been studied using a series of SO2 complexes of palladium and platinum and the results contrasted with adducts of SO2 with main group Lewis acids and bases. The isomers {(PH3)2M(SO2) p}16 and {(PH3)2M(SO2) np}14 are calculated to have very similar energies and the relative stabilities of analogous isomers may be manipulated by varying the bite angle of the phosphine ligands in {[(PH2)2CnH2n]M(SO2)}. Keywords Linear nitrosyls, non-linear nitrosyls, bent nitrosyls, sulfur dioxide, iodine, nitrido, imido ambiphilic, ambivalent, amphoteric valence tautomerism, DFT calculations, gasotransmitters, catalytic reactions of nitrosyls, trans- influence.

1

Page 1 of 34

1 Introduction

Ac

ce pt

ed

M

an

us

cr

ip t

The initial extension of the Lewis bond description to Werner co-ordination compounds by Sidgwick was based on the assumption that the ligand is an electron pair donor to the Lewis acidic metal cation. As co-ordination chemistry has developed it has become increasingly clear that this description of the metal-ligand bond is an oversimplification. Ligands may enter into multiple bond interactions and indeed metals may also donate electron pairs to Lewis acids [1-5]. Pauling [6] was the first to articulate the idea that ligands such as CO are able to function simultaneously as a Lewis acid and Lewis base, when co-ordinated to transition metals in low oxidation states. These synergic interactions involve complementary σ and π orbitals on the ligand and metal and the point group symmetry of the adduct remains essentially unaffected by the relative contributions of the forward and back donation components, although the relative lengths of the M-C and C-O bonds do change [1]. The formal two-electron donating abilities of ligands such as CO remains constant from complex to complex, but variations in the forward and back donation components can change the electron distribution in the metal-ligand moiety and influence the reactivities of the complexes with nucleophiles and electrophiles. Changes in the oxidation state, co-ordination number, steric effects and the donor/acceptor properties of the “spectator” ligands have been used to fine-tune these reactivity trends. In a recent review [7] I have drawn attention to the common characteristics of ligands which are capable of adopting alternative geometries, because of their ability to vary the number of electrons which they formally donate to the metal. The differences between ambidentate, ambivalent and ambiphilic ligands have been clarified and a notation has been developed which may be used to define the alternative geometries they adopt when co-ordinated to transition metals. A summary of this notation is provided in the Appendix and the reader is directed to reference [7] for a fuller description of the proposals and specific examples of its applications. The notation is equally applicable to ligands which are π-acceptors and π-donors. This similarity between π-donor and acceptor ambivalent ligands has been understated previously. The ambivalent character of other πacceptor and π- donor ligands is summarized in Table 1 and illustrated in Figure 1. The alternative geometries adopted by these ligands may be related in the first instance to the electron donating and accepting abilities of the ligands within the traditional valence bond framework as shown in Figure 1. Whilst this methodology, when used with the 18 electron rule, provides a satisfactory mode of accounting qualitatively for the alternative geometries of the majority of complexes, it fails to provide a satisfactory description of complexes with intermediate geometries, the relative energies of the isomers with the alternative geometric forms, the effects of the alternative geometries on the other metal-ligand bonds and the consequences on the reactivities of the complex. The aim of this review is to provide a deeper insight into the bonding properties of ambivalent and ambiphilic ligands within a molecular orbital framework. The DFT methodology used for this theoretical analysis is described in more detail in the Appendix. Although this review aims to provide a more detailed understanding of the broad class of ligands which exhibit ambivalent and ambiphilic properties it is hoped that the conclusions derived from the analysis may also have implications for understanding the roles of such ligands in catalytic and biological processes. The discovery that nitric oxide has many roles in biology has resulted in an exponential growth of research into its chemistry and biochemistry. This has included detailed studies of a wide range of nitrosyl transition metal complexes of iron and copper with ligands which resemble those found in biology [8,9]. It has also renewed interest in the co-ordination chemistries of a range of the related SO2, H2S, CO, COS and N2O molecules, which may also function as neurotransmitter molecules in biology [8-10]. When these gasotransmitters co-ordinate to transition metals they trigger subtle, but important, changes in the effective size of the metal (via spin changes), changes in 2

Page 2 of 34

M

an

us

cr

ip t

Table 1. Summary of the electron donating capabilities of ambivalent ligands with symmetry signatures Ligand M-L geometric Electron description donation (descriptor) NO,NS Linear (l)(180-160o) 3 electrons NO,NS Bent (b)(100-140o) 1 electron NR2(PR2) Non-planar (np) 1 electron NR2(PR2) Planar (p) 3 electrons NCR2 (PCR2) Linear (l) 3 electrons NCR2(PCR2) Bent (b) 1 electron N2R Linear (l) or Singly 3 electron bent (sb) N2R Doubly bent (db) 1 electrons N2R π-bonded (η2) 3 electrons NR Linear (l) 4 electrons NR Bent (b) 2 electrons NOR Linear (l) 4-electrons NOR Bent (b) 2 electrons OR Bent (b) 1 or 3 electrons OR Linear (l) 5 electrons Table 2. Summary of the electron donating capabilities of ambiphilic ligands and metal complexes

ce pt

With symmetry signatures SO2 SO2 I2 I2 IrCl(CO)(PPh3)2 IrCl(CO)(PPh3)2

Ac

M-L geometric description (descriptor)

Electron donation

Planar (p) Non-planar (np) Bent (b) Linear (l) Angular C2v Square planar

2 electrons 0 electron 2 electrons 0 electron 0 electron 2 electrons

ed

Ligand

the protein environment, via trans- influence effects, and the modulation of the redox properties of the metallo-protein. Understanding in molecular terms the mode of action and selectivity of the interactions between nitric oxide, dioxygen, etc and metalloproteins is fundamental. Specifically it is important to understand those factors which enable ligands to alter their geometries on coordination to transition metals and the impact on the conformations and subsequent reactions of metalloproteins. Basolo and his co-workers’ [11] seminal mechanistic studies on complexes, which contain ligands capable of varying the number of electrons which they donate to the metal, established that they proceed by a different mechanism to that established previously for carbonyl complexes. They proposed that the transfer of an electron pair to the ambivalent ligand was associated with the bending process and opens up an empty orbital on the metal and thereby facilitates a bimolecular rnucleophilic substitution reaction at the metal. In recent years Berke and his co-workers [12] have established the

3

Page 3 of 34

ip t

catalytic implications of this proposal for interpreting and enhancing the catalytic reactions of metal nitrosyl complexes. The description of the properties of these ligands has clear implications for understanding their catalytic and biological effects. It is therefore necessary to provide clear definitions of the different classes of ligands and understand their structural and electronic properties at a molecular level. Although specific detailed molecular orbital calculations have been reported for a wide range of ambiphilic and ambidentate ligands and used to interpret their geometries and specific aspects of their reactions [13-35] there has been no attempt to understand in more general terms their electronic and structural features.

us

cr

An ambidentate ligand has two or more Lewis base sites with potential donor capabilities. Generally these are lone pairs on alternative donor atoms, e.g. SCN- or NCS-, but one of the isomers may involve donation from a π or σ bond, e.g. O2 or H2. Donation of electron pairs from these alternative sites may lead to different atom sequences in the complex and different symmetries, although in each isomer the EAN count is identical [1,7].

ed

M

an

An ambivalent ligand is capable of forming more than one bond to a transition metal by donating a variable number of electrons. In these complexes the initial donor-acceptor bond is supplemented by donation from lone pairs on the donor atom to empty orbitals on the metal. These multiple interactions lead to multiple metal-ligand bonds and the dative π- component is enhanced by adopting a higher symmetry geometry (linear or planar usually). These ambivalent ligands are therefore associated with a symmetry signature, which is defined below. In the Green-Parkin scheme [1] for organometallics this corresponds to a change from bent NO acting as an X ligand to an LX ligand in a linear NO complex. In π-bonded organometallic complexes the carbocyclic ligands may also give rise to pairs of compounds which result from a reduction in the metal-ligand hapticity by 2. The additional electron pair resides in a disengaged π orbital on the carbocyclic ligand rather than in a lone pair orbital [7].

ce pt

An ambiphilic ligand has a HOMO and LUMO which are both readily accessible to a Lewis acid or base. If the HOMO and LUMO have different symmetries then this may be reflected in alternative geometries of the Lewis base and acid adducts, e.g. planar and non-planar SO2. In the Green-Parkin scheme planar SO2 behaves as an L (2 electron donor) and non-planar SO2 acts as a Z (0 electron) acceptor ligand [1] (see Table 2).

Ac

I have introduced the description symmetry signature [7] to convey the impression that the donation of an electron pair from the localized lone pair orbital on the ligand to the metal (see Figure 1) results in a change of symmetry of the metal-ligand moiety, e.g. the M-X-Y bond angle for a metal complex between the ligand XY and the metal M increases from approximately 120 o, and in the limit approaches 180o. However, the variation in M-X-Y bond angles is greater than that observed in organic and main group molecules because the electronic properties of the metal may be influenced by the donating characteristics and geometries of the other ligands co-ordinated to the metal and the metal oxidation state. The ambivalent and ambiphilic ligands share the pairs of structures listed in Tables 1 and 2. A lower symmetry structure, which has a lone pair localized on the ligand, and a higher symmetry structure, which involves the lone pair donating to an empty orbital on the metal, is the common characteristic. The stereochemical activity of the lone pair leads to an angular (M-X-Y) or non-planar (M-XY2) structure in accordance with VSEPR theory. Donation of this lone pair to an empty orbital on the metal leads to an increase in multiple bond character and the adoption of the higher symmetry structure (linear or planar). Some specific examples of ligands showing symmetry 4

Page 4 of 34

ip t cr us an M ed ce pt

Figure 1. Examples of ambidentate, ambivalent and ambiphilic ligands and their formal electron donating abilities.

Ac

signatures are illustrated in Figure 2 [7 ]. Figure 3 illustrates in qualitative molecular orbital terms the interaction between the filled lone pair orbital and the empty metal orbitals and emphasizes the electronic similarity between NR and NO as ligands. The effective transfer of an electron pair from N to the metal suggested by Figure 2 depends not only on the electronegativity difference and the overlap integrals between the ligand and the metal, but also on the “spectator” ligands. This molecular orbital description provides a degree of flexibility which is absent with the valence bond framework since it permits a variation in the extent of donation from the lone pair to the empty orbital on the metal and therefore incorporates the possibility of intermediate geometries. The most extensive data for linear (l), bent (b) and intermediate (i) structures are available for nitrosyl complexes. A recent summary and statistical analysis of the structures of more than 2700 nitrosyl complexes is given in reference 8. More than 90% are classified as linear {LmM(NO)n l}y (M-N-O 180-160o) and 4% have bent {LmM(NO)n b}y (M-N-O 110-140o ) and there are 4% with intermediate {LmM(NO)n i}y geometries (M-N-O 140-160o ).[8] Related statistical analyses of other ambivalent ligands would provide a useful structural database for these ligands. A qualitative overview of the relevant structural data suggests that the nitrosyl distribution is not necessarily typical. 5

Page 5 of 34

ce pt

ed

M

an

us

cr

ip t

Table 3 summarises the calculated Ru-N-O bond angles in a series of square-pyramidal complexes using the DFT methodology. The results underline the important point that even in a series of isostructural and isoelectronic complexes the M-N-O bond is influenced by the donating characteristics of the X basal ligand. The M-N-O bond angles vary between 129 and 159o and do not separate neatly into the sp (180o) and sp2 (120o) descriptions familiar to organic chemists. Specifically the stronger π-donor ligands favour an increase in the M-N-O bond angle. A molecular orbital interpretation of this phenomenon was presented by Hoffmann, Mingos and their coworkers [13] and emphasised the limitations of simpler VSEPR and hybridisation analyses.

Ac

Figure 2. Some specific examples of complexes of ambivalent ligands [7,8].

Figure 3. Qualitative molecular orbital description of bonding in complexes of ambivalent NO and NR ligands (with symmetry signatures).

Table 3. Summary of calculated bond angles [RuCl(NO)(X)(PH3)2]+ as a function of the ligand X. 6

Page 6 of 34

π-acceptor π-acceptor π-acceptor π-acceptor π-acceptor π-acceptor π-acceptor π-donor π-donor

cr

ip t

α(o) 129 130 132 143 139 136 161 159

X NO NS N2H PO NCH2 SO2 NH2 PH2

Ac

ce pt

ed

M

an

us

It is also worth emphasising that some ligands such as alkynes are capable of donating alternative numbers of electrons without leaving a symmetry signature, because the donation occurs from a πbond perpendicular to the R-C-C-R plane rather than a lone pair orbital and therefore no change in symmetry is required (see Figure 4). Such ligands are however relatively rare. Strong circumstantial evidence from bond length data has been accumulated to justify the proposal for single or double πdonation in complexes of alkyne ligands [33]. The purpose of this review therefore is to amplify the broadly based system of classification summarised in Tables 1 and 2 with molecular orbital calculations based on the DFT approximation (see Appendix for more detailed description of the DFT calculations [36,37]). Although there have been many papers and reviews [14-34] providing a detailed analysis of the bonding in specific complexes of ambivalent and ambiphilic ligands this review attempts to consider some broader general issues. It is perhaps a reflection of how modern inorganic chemistry is pursued that the majority of the previous research has concentrated on developing an understanding of a particular ligand and almost no work has been done on complexes with combinations of ambivalent ligands and data on combinations of π-acceptor and π-donor ambivalent ligands within one complex is virtually non-existent [34]. It is hoped that by drawing attention to ambivalent ligands as a whole and providing some theoretical predictions more experimental work in the area will be encouraged. In the absence of experimental data the DFT calculations have been used to highlight possible interesting broad trends. Specifically the following topics have been addressed: 1. The energetics for the transformation between the alternative linear – bent or planar-nonplanar geometries have been estimated. 2. In complexes where more than one ambivalent ligand is present theoretical calculations have been performed to place the ligands in an order which reflects their ability to form bent or non-planar structures. 3. The relative trans- influences of ambivalent ligands in their two structural forms have been estimated, because of the importance of trans- labilization by ligands such as NO in biology and catalytic processes. 4. The valence tautomeric fluxional processes involving ambivalent and ambiphilic ligands has been studied and the calculated energies have been compared with the available experimental evidence. 5. Complexes with more than one ambidentate ligand may either adopt geometries with some of the ligands linear and others bent, e.g. [OsO2(NBut)2], or more symmetrical structures with all the ligands showing intermediate geometries, e.g. [Os(NBut)4]. The 7

Page 7 of 34

cr

ip t

electronic factors which lead to this difference in behaviour are not well understood and a detailed theoretical study and analysis may shed some light on this subtle problem [2732].

ce pt

ed

M

an

us

Figure 4. Dihapto- alkynes are able to function as ambivalent ligands which do not display symmetry signatures.

Ac

Figure 5. Experimental [38-40] and theoretically calculated structural data for two complexes, each of which has a pair of symmetry inequivalent ambivalent ligands. The notation used to describe these complexes has been described in more detail in the Appendix and references 7 and 8.

2. Results of DFT Calculations [34-36] 2.1 Geometric features Figure 5 compares the important structural parameters for a pair of nitrosyl and dialkylphosphidocomplexes and the results of DFT theoretical calculations on idealised complexes [38-41]. These complexes were chosen, because they both contain within one complex the same ambivalent ligand in two distinct co-ordination environments, viz linear and bent nitrosyl and planar and non-planar phosphido-. In the ruthenium complex the ambivalent ligands occupy axial and basal sites of a square-pyramid and the alternative geometries are associated with the symmetry inequivalent sites. 8

Page 8 of 34

Ac

ce pt

ed

M

an

us

cr

ip t

For the hafnium bent sandwich complex the dicyclohexylphosphido- ligands co-ordinate in the plane perpendicular to that involving the tilted cyclopentadienyl- ligands. In contrast to the square-pyramid the alternative phosphido- geometries are observed although they occupy symmetry equivalent coordination sites. The bonding in complexes of ambivalent ligands was first analysed using the Walsh diagram molecular orbital methodology as early as 1973[13] for NO and subsequently for the other ligands listed in the Table 1 [14-32] using more sophisticated calculations. In recent years it has become more routine to undertake DFT calculations on organometallic and co-ordination complexes either based on model complexes or using combined DFT and molecular mechanics calculations and some recent reviews are to be found in references 35 and 36. The approach we have used is described in the Appendix, but briefly we have chosen to use model compounds based on PH3 rather than PPh3 or PCy3 and PMe2 for PCy2 since we were more concerned with exploring general trends rather than analysing a specific feature in great detail. The energy minimised DFT calculations for the model compounds (summarised in Figure 5) reproduce the local geometries well; with the observed individual bond lengths agreeing to within 0.08Å and the angles to within 7o. The geometries around the metal atom (square pyramidal around ruthenium and tetrahedral around hafnium if the centres of the rings are used to define the centroids of the Cp ligands) and the presence of the distinct two symmetry signatures for NO and PR2 respectively accord with the assigned structural descriptors. The bending of the NO towards the linear NO rather than the trans- Cl in the ruthenium compound is also reproduced. Figure 3 has illustrated the qualitative molecular orbital ideas, which may be used to account for the shorter metal-ligand bond lengths for the ligand with the more symmetric geometry. The greater metal-ligand multiple bond character in the more symmetrical isomer and the greater scharacter in the metal-ligand σ-donor bond both contribute to a M-L shorter bond. The DFT calculations reproduce the shorter bond lengths in the {M(NO) l} and {M(PR2) p} fragments reasonably well and the agreement between experimental and calculated data is particularly good for the phosphido- ligands (see Figure 5 and Appendix). The study of complexes with two identical ambivalent or ambiphilic ligands immediately raises two interesting questions: firstly do the ligands adopt the alternative extreme geometries summarised in Figure 1 or do they adopt a more symmetric geometry with each ligand adopting an intermediate geometry and secondly what are the detailed differences in metric parameters if the ligands are symmetry inequivalent? The DFT calculations on model [RuCl(NO)2(PH3)2]+ have confirmed that the square-pyramidal structure with an equatorial linear nitrosyl and a bent apical nitrosyl is 6.1 kcalmol-1 more stable than an isomer with a trigonal pyramidal structure and a pair of equatorial nitrosyls with slightly bent nitrosyls. The corresponding calculations on + [OsCl(NO)2(PH3)2] suggests that the isomer with both nitrosyls essentially linear is stabilised relative to the linear/bent isomer and is only 0.1 kcalmol-1 less stable, i.e. the two isomers are effectively equienergetic. The symmetrical isomer is associated with a local geometry around the metal which resembles a trigonal-bipyramid more closely. The similarity in energies of the two isomeric forms is not surprising since the interconversion process is similar to the Berry pseudo- rotation, although complicated by the simultaneous interconversion of linear and bent nitrosyls. These results on model compounds suggest that these isomeric forms of dinitrosyl compounds have very similar energies and the energy difference is sensitive to the central metal atom even when the metal atoms belong to the same group of the periodic table. This observation is consistent with the reports that although [RuCl(NO)2(PPh3)2]+ has the geometry illustrated in Figure 5, the closely related [OsCl(NO)2(PPh3)2]+ complex has a trigonal-bipyramidal structure with an attracto- geometry (Os-NO = 170o) [47]. The bond angles for the non-phosphine ligands in the closely related ruthenium and osmium complexes are shown below in 1 [38,41-43]. The model calculations reproduce the

9

Page 9 of 34

ip t

stabilisation of the trigonal-bipyramidal isomer for osmium but underestimates its stabilisation. More detailed

ed

M

an

us

cr

1 calculations on the very soft potential energy surface using a combination of DFT and molecular mechanics calculations would perhaps give a more satisfactorily account for the effects of the PPh3 ligands. The structural data on [OsCl(NO)2(PPh3)2]+ also suggests that replacing the Cl ligand by OH stabilises the square-pyramidal isomer suggesting that the ligand changes also result in dramatic structural effects . These structural and theoretical data confirm that energy differences between symmetric and asymmetric isomers of dinitrosyl complexes, [MX(NO)2(PR3)2]+ , are small (less than 10 kcalmol-1) and the relative stabilities of the two isomeric forms may be influenced by small changes in the ligands and the central metal atom. The presence of symmetry distinct sites in a coordination polyhedron may also contribute to the presence of linear and bent geometries within one complex. The preference of the apical NO in a square-pyramidal complex to bend is well established and was interpreted within a molecular orbital framework 40 years ago [13]. These observations reinforce the view that metal complexes with two ambivalent ligands may have symmetric and asymmetric geometries which have very similar energies. In the recently introduced notation [7,8] they correspond to {L3M(NO)2 l,b}16 and {L3M(NO)2 l,l}18 and reflect the transition metal’s ability to accommodate formal 16 and 18 electron counts [54].

Ac

ce pt

2.2 Valence tautomerism –experimental evidence When a pair of ambivalent/ambiphilic ligands adopt the alternative metal-ligand geometries within one complex the intramolecular interchange of the ligands constitutes a valence tautomerism. The formal electron pair movements associated with such a valence tautomerism are illustrated in a general way in Figure 6 for complexes of ambivalent/ambiphilic XY and XY2 ligands. Interestingly although valence tautomerism is relatively common in organic and organometallic chemistry it is rarely observed in co-ordination complexes. The calculations described above on the model ruthenium dinitrosyl complexes suggests that the energy surface connecting the tautomers is rather soft, especially for simple diatomic ligands, but may be larger for more complex ligands where steric effects may also contribute to the activation barriers. The square-pyramidal [MX(NO)2(PPh3)2]+ (M =Ru or Os, X = Cl or OH) complexes have been more closely studied experimentally. Circumstantial evidence for an intramolecular tautomerism comes from the structural studies of closely related complexes and infrared studies on the complexes in the solid and solution states and their labelled isotopomers. [38,41-45]

10

Page 10 of 34

ip t cr us an

Figure 6. Valence tautomerism in complexes with a pair of ambivalent or ambiphilic ligands.

Ac

ce pt

ed

M

Circumstantial evidence for the soft potential energy surface comes from the infrared spectra of solutions of the ruthenium and osmium complexes which show additional bands compared to those observed in the solid state. This and associated studies on the labelled analogues have suggested that not only do the nitrosyls interconvert very rapidly, but also more than one isomer exists in solution [41-50]. This introduces additional peaks and troughs into what is already a quite soft potential energy surface and places the authoritative analysis at the limit of current DFT methods for metal complexes of the heavier transition metals. Solid state 15N nmr studies have demonstrated that the nitrosyls in the complex 15 [RuCl( NO)2(PPh3)2]+ do not interchange in the solid state [48]. The two distinct nitrosyl ligands are easily distinguished by the large chemical shift difference (several hundred ppm) and also the anisotropies in the chemical shift tensors of the axially symmetric and the bent NO ligands. In solution the nitrosyl ligands in this complex do indeed undergo a rapid intramolecular fluxional process which makes the nitrosyl environments equivalent, but it must be associated with a small activation energy since little line broadening is observed when the temperature is lowered. The nitrosyl ligands are made equivalent by an intramolecular process which does not involve dissociation of the NO+ or phosphine ligands [41-50]. The presence of isomers in solution and their rapid interconversion on an nmr time scale were confirmed by the equilibrium isotope effects observed in the variable temperature solution 15N{1H} nmr studies on [RuCl(NO)(15NO)(PPh3)2]+ and [OsCl(NO)(15NO)(PPh3)2]+ [46-51]. Figure 7 illustrates the geometric consequences for the intramolecular interconversion of these dinitrosyls of ruthenium and osmium and distinguishes tautomeric and isomeric processes. The interconversion shown at the top of Figure 7 does not lead to a valence tautomerism but interconverts the original nitrosyl, isomer 1, to a second, isomer 2, with the nitrosyl bent towards chloride rather than the linear nitrosyl - such a process would not interconvert the linear and bent isomers since the linear and bent forms retain their original identities. The nitrosyl interchange may not necessarily proceed by the increase of the M-N-O from 120o to 240o., but may proceed via a trigonal-bipyramidal isomer (3r) which has both nitrosyls in symmetry equivalent locations. The interconversion shown in 11

Page 11 of 34

an

us

cr

ip t

Figure 7 involving the simultaneous angular changes of both nitrosyls does lead to an interchange of the nitrosyls and therefore represents a tautomerism rather than an isomerisation. The simultaneous distortions of both nitrosyls may involve a symmetrical trigonal-bipyramidal isomer (3a).

ed

M

Isomer 1 2 3a,3r + [RuCl(NO)2(PPh3)2] 0 4.6 6.1 kcal mol-1 [OsCl(NO)2(PPh3)2]+ 0 9.0 1.0 kcal mol-1 Figure 7. The illustration of the atomic motions which lead to the interconversion of the isomers of [MX(NO)2(PH3)2]+ and the relative energies of the isomers.

Ac

ce pt

The DFT calculations, summarised in Figure 7, confirm that the alternative isomeric forms have very similar energies and that for the ruthenium complex the isomer 1 is more stable than isomer 2 by 4.5 kcal mol-1 and the repulso- and attracto- isomers 3a and 3r by approximately 6 kcal mol-1. The energy difference between isomer 1 and isomer 3a for the corresponding osmium complex is reduced to 1 kcal mol-1 and increased between isomers 1 and 2 to 9.0 kcal mol-1. This is consistent with the structural data which established that in the solid state [RuCl(NO) 2(PPh3)2]+ has the structure 1 and [OsCl(NO)2(PPh3)2]+ has the structure 3a . In the latter example the nitrosyls make an M-N-O angle of 170o and adopt an attracto- geometry, rather than a repulso- geometry with the nitrosyls bending away from each other. These DFT calculations therefore reproduce the geometric preferences for d8 5 co-ordinate nitrosyl complexes reasonably well and confirm earlier proposals based on perturbation theory arguments and semi-empirical molecular orbital calculations [13]. The presence of four isomeric structures differing in total energies by only 6.2 kcalmol -1 within a narrow area of reaction space ensures a soft potential energy surface and the transition states for the interconversion on such a soft potential and complex energy surface are difficult to verify unambiguously. However, little doubt remains that the energies separating structures with linear/linear and linear/bent nirosyls is small. Alvarez and Liunell [51] have completed a thorough Dunitz-Bürgi type analysis of the structures of five co-ordinate nitrosyl complexes and shown that they fall into two distinct categories based on either the trigonal-bipyramid and square- pyramid [52]. The distribution of experimental data suggest that there are two minima on the potential energy surface – one based on a square12

Page 12 of 34

us

cr

ip t

pyramid with a severely bent nitrosyl and the other on a trigonal-bipyramid with approximately linear nitrosyl. The scarcity of experimental structures with intermediate geometries suggests a barrier to interconversion and clear evidence for distinct symmetry signatures for these complexes. Therefore, complexes of this type provide examples of double symmetry signatures, which are observed at the ambiphilic ligand and at the ambiphilic metal sites [7]. Following the designations given in Tables 1 and 2 for ligands the changes in metal geometry for the ambiphilic metal complex may also be indicated as follows:{sp L3M(NO)2 lb}16 and {tbp L3M(NO)2 ll}18 , where sp and tbp represent the square pyramidal and trigonal-bipyramidal metal geometries.

Figure 8 . Valence tautomerism in a hafnium di(phosphido-) complex.

ce pt

ed

M

an

The diphosphido- complexes of hafnium [Hf(η-C5H5)2(PR2)2] (see Figure 8) has both planar and non-planar phosphido- ligands ( R =Et) in the solid state. For R = Cy two 31P signals are observed at low temperatures and they coalesce at higher temperatures indicting a valence tautomerism similar to that described above for nitrosyl complexes. The computed activation energy is approximately 6 kcal mol-1. Analogous arsenido- complexes are known and a molybdenum complex undergoes a similar planar-non-planar valence tautomerism [39,40]. Valence tautomerism has also been observed in alkyl imido- complexes which have linear and bent geometries within the same molecule, e.g. [OsO2(NBut)2] and the molecules are fluxional on the nmr time scale with an activation energy less than 5kcal mol-1 [53-54]. Unfortunately, sulphur dioxide complexes are not so amenable to such studies because the sulphur and oxygen nuclei are not as favourable for nmr studies [55-60]. These limited studies available indicate that the activation energy for the tautomerism involving the isomeric forms of ambivalent and ambiphilic ligands is 5 kcal mol-1 or less.

Ac

2.3 Relative bending abilities of ambivalent ligands The presence within the same complex of two different ambivalent ligands raises the question: “which ligand is more likely to adopt a bent (or non-planar geometry)”. Experimentally this problem has not been studied systematically although there is some structural data providing circumstantial evidence for example that SO2 adopts a non-planar geometry more readily than NO adopts a bent geometry [55-60]. There is also an absence of thermodynamic data on the relative energies of the isomeric forms of ambivalent complexes. To establish preliminary data on these structural issues the energies of the isomeric square-pyramidal complexes analogous to those shown in Figure 5 were calculated and the results of the calculations are discussed below. These complexes provide a convenient test-bed for assessing computationally the relative abilities of ambivalent ligands to adopt a bent rather than a linear geometry relative to NO or PR2. The square-pyramidal ruthenium complexes provided a more reliable series of complexes for this investigation because it retains the integrity of the basic structure, whereas for the hafnium complexes ligand rearrangements complicate the analysis. For example, for the hafnium bent sandwich compounds the replacement of one of the PR2 ligands by NO leading to [Hf(η-C5H5)2(NO)(PMe2)] [39,40] provides in principle an opportunity of comparing two related complexes with either NO bent

13

Page 13 of 34

or PMe2 non-planar. However, the energy minimisation led to a preference for NO adopting a η2- and a PMe2 non-planar geometry. The study of the corresponding di(nitrosyl) hafnium complex led to the

cr

ΔE (kcal mol-1) 0 6.85 -13.96 -0.59 5.64 -0.40 -10.30 1.60

us

Xbasal NO NS PO N2H NCH2 SO2 PH2 NH2

an

Xaxial NO NS PO N2H NCH2 SO2 PH2 NH2

ip t

Table 4 Comparison of the relative stabilities for complexes of isomeric ambivalent ligands

Ac

ce pt

ed

M

formation of a hyponitrito- complex resulting from the formation of an N-N bond between the nitrosyls. The square-pyramidal ruthenium complexes were better behaved in this respect and were therefore used to study the geometric preferences. The DFT calculated results for a range of ambivalent ligands are summarised in Table 4 . The calculations are referenced with respect to the structurally characterised ruthenium dinitrosyl complex for which ΔE is equal to 0. A negative value of ΔE suggests that the ligands favours bending relative to NO. For this series of complexes the ambivalent and ambiphilic ligands follow the following order of bending abilities: PO > PH2 > N2H > SO2 > NO > NH2 > NS. The electronegativity of the donor atom plays a crucial role in influencing the relative stabilities of the isomers. For example PO and PH2 prefer the ligand to adopt the less symmetrical geometry than NO and NH2. This is consistent with the initial molecular orbital analyses [13,54,61], which relates the steepness of the reaction co-ordinate for the bending process to the effectiveness of the mixing between π*(NO) and dz2 . The introduction of a less electronegative substituent on the diatomic ligand favours the bending process because ligand based component lies at a lower energy. In contrast the introduction of a less electronegative substituent in the non-donor location disfavours the bending process, e.g. NS vs NO. The qualitative bonding model illustrated in Figure 3 also provides a basis for understanding these differences since donation of a lone pair on the ambivalent ligand is energetically favoured if the donor atom belongs to the second long series of the periodic table, because of the lower electronegativity of the donor atom. For the ligands studied the largest difference in energy approaches 15 kcalmol-1 and the least favoured is NS which is disfavoured relative to NO by approximately 7 kcalmol-1. It is significant that the energetics of NO and N2H linear and bent geometries are very similar and that the adoption of a non-planar geometry by the SO2 ligand is favoured relative to NO. This is consistent with the X-ray data reported by Kubas et al [5560]. The ligands which have a greater preference for the bent geometry, e.g. PH2 and PO, are more likely to form asymmetric diligand complexes, which have different ligand geometries, rather than a more symmetric structure with equivalent ligands. The structure for [Hf(C5H5)2(PR2)2] shown in 14

Page 14 of 34

Figures 5 and 8 provides a specific example of such a complex. The reliability of the data in Table 4 and specifically the relative ability of ligands to adopt the alternative geometries should be tested experimentally using a combination of structural, infrared and variable temperature nmr studies.

Ac

ce pt

ed

M

an

us

cr

ip t

2.3 Trans-influences in complexes with ambivalent ligands Jiang and Berke [12] have noted the importance of the trans- influence of the bent nitrosyl ligand in the catalysis of organic transformations using nitrosyl complexes. The biological role of NO is also intimately connected with the effect it has on the trans- ligand in an octahedral iron(II) porphinatocomplex and therefore if a wider range of gasotransmitters are to be studied in future it is important to establish the relative abilities of ambivalent and ambiphilic complexes to exert a significant transinfluence. As early as 1973 I established that the trans- influence of NO was intimately connected with the bending of the NO ligand [61]. These calculations were based on a Walsh diagram analysis of the frontier molecular orbitals of [Co(NO)(NH3)5]2+ using the Wolfsberg-Helmholz approximation. In essence the bending process involves the mixing of π*(NO) and the metal dz2 orbitals and since the latter is also strongly antibonding to the trans- ligand in an octahedral complex the bending distortion causes a significant lengthening of the trans- Co-N bond. This section reports recent DFT calculations which compare the relative trans- influences of ambivalent and ambiphilic ligands. Some years ago we compared the trans- influences of the linear nitrosyl and the nitrido- ligand in related 18 electron octahedral ruthenium complexes and the results have been recalculated and are summarised in Figure 9 [25,27]. The shorter calculated Ru-N bond lengths to the nitrido and nitrosyl ligands in [Ru(X)(NH3)5]3+ (X = N or NO) compared with the Ru-(NH3)cis- bonds confirm the long held view that both NO and N form multiple bonds to ruthenium. [1,4,25,27]. The calculations also suggest that the multiple bonding is stronger for the Ru-N (nitrido-) bond and this has been related to the orbital responsible for multiple bonding being more localised on the nitrogen of the nitrido- ligand than the nitrogen of the nitrosyl ligand. The calculations highlight an important difference between the nitrido- and linear nitrosyl ligand. Although they are both formally 3 electron donors which are capable of multiple bonding the trans- Ru-NH3 bond is lengthened significantly in the nitridocomplexes, i.e. by 0.33Å in [RuN(NH3)5]3+ and by 0.12Å in [RuNCl5]2- . The electronic reasons for this difference have been analysed in some detail by Lyne and Mingos [25,27], who related the stronger trans- influences in the nitrido- complexes to the stronger multiple bonding in these complexes. This is enhanced by a bending of the equatorial chlorine ligands away from the strong RuN bond and this results in a weakening of the trans- Ru-Cl bond. Mountford and Kaltsoyanis analysed the trans- influence of the alkylimido- ligand and came to similar conclusions [17].

Figure 9. Summary of calculated bond lengths in nitrosyl and nitrido- octahedral complexes of ruthenium.

15

Page 15 of 34

Ac

ce pt

ed

M

an

us

cr

ip t

Table 5 summarises the calculated Ru-Cl bond lengths in a series of 18 electron octahedral {Ru(XY)Cl5 l}y- and {Ru(XY2)Cl5 p}y- complexes (XY = NO, NS, PO, N2, NH, N2H b); XY2 = SO2 ), which support the view that the trans- influences in this series of complexes with the more symmetrical structures, i.e. linear for XY and or planar for XY2 do not have very significant transinfluences. Indeed the ambivalent ligands may even show a negative trans- influence, i.e there is a slight shortening of the trans- metal-chlorine bond, but a more thorough analysis would be required to verify the statistical significance of the observed differences. The only ambivalent ligand which shows a small positive trans- influence is SO2. This contrasts with the large trans- influence noted previously for the nitrido- ligand in [RuNCl5]2-. In the less symmetrical 18 electron complexes {Pd(XY)Cl5 b}y- and {Pd(XY2)Cl5 np}y- which have bent M-X-Y and non-planar M-XY2 geometries, a large trans- influence is suggested by the calculations. In the 1970s it was shown that the mixing of orbitals in nitrosyl complexes as the M-N-O bond angle changes from 180 to 120o contributes to the weakening of the trans- metal- ligand bond [61]. The occupied dz2 orbital has a substantial antibonding contribution involving the trans- metalligand bond and mixes with π*(NO-dxz,yz). The chemical and biological implications of transinfluence effects for NO are significant and therefore it is important to establish whether other ambivalent gasotransmitter ligands exhibit similar characteristics. Table 6 summarises the transinfluence for a series of octahedral palladium complexes {Pd(XY)Cl5 b}y- and {Pd(XY2)Cl5 np}y- with ambivalent ligands which have the ligand co-ordinated in the less symmetrical manner. It is noteworthy that although NO and PH2 ligands have the largest trans- influences in the series studied the other ambivalent ligands also exhibit significant trans- influences. It goes without saying that a trans- influence is a ground state structural phenomenon, but in general it is reflected in a strong trans- effect whereby the trans- ligand is labilized in nucleophilic reactions of the parent complex. This phenomenon has certainly been observed in nitrosyl complexes and it is probably also responsible for the enhanced labilities of the amine Co(III) complexes towards nucleophilic substitution reactions under basic conditions originates from the high labilizing effect of the NH2 ligand [62]. The calculated structural data in Tables 6 and 7 underline the important general point that the trans- influences of ambivalent and ambiphilic ligands are much greater in those complexes which have the less symmetrical metal-ligand geometry. The two additional valence electrons introduced in the palladium complexes occupy a molecular orbital with additional antibonding character between the metal and the trans- ligand and this is responsible for the larger trans- influences in this series of complexes. This relationship between the geometry of the co-ordinated NO ligand and its trans- influence in more biologically relevant molecules has been demonstrated by structural studies of the metalloporphyrin complexes [MII(TPP)(L)(NO)] (L = 4-methylpiperidine) (M = Mn, Fe and Co) [61,63-69]. For M = MnII ({L5M(NO) l}18 ), not only is the Mn-NO angle nearly linear (176o), the Mn-Npip bond length is relatively short (2.20 Å). For M = FeII ({L5M(NO) i}19 ), the Fe-NO angle is bent to 142o, and the bond to the methylpiperidine is considerably weakened (Fe-Npip = 2.46 Å) . For M = CoII, ({L5M(NO) b}18 ), the Co-NO angle is even sharper and the trans- influence is greater, and a stable complex with methylpiperidine could not be isolated. It was on the basis of this trans- influence for the {L5M(NO) i}19 case that Traylor and Sharma proposed a mechanism for sGC activation by NO [63]. Soluble guanylyl cyclase (sGC) is a haem enzyme with a FeII(PPIX) ("hemin", PPIX = protoporphyrin IX) as the metal centre with an open axial coordination site (the distal site) . The proximal site is occupied by a histidine nitrogen. Coordination of NO to the haem centre gives the {[L5Fe(NO) i}19 ), complex and the associated transinfluence weakens the proximal histidine-iron bond leading to changes in protein conformation which activates the enzyme by several orders of magnitude. An impressive test of this proposal was offered by Burstyn and coworkers [68], who investigated the activities of non-native sGC prepared by 16

Page 16 of 34

ip t

substituting MnII(PPIX) and CoII(PPIX) for the hemin of the native enzyme [68]. Addition of NO failed to activate sGC(Mn) above basal activity, presumably because the proximal histidine was not labilized in this ({L5Mn(NO) l}18 ), complex. In contrast, NO addition to sGC(Co) giving a ({L5Co(NO) b}18 ), complex resulted in even greater activity than with sGC that had been reconstituted with hemin. The overall trend, sGC(Co)(NO) > sGC(Fe)(NO) >> sGC(Mn)(NO) substantiates the Traylor/Sharma hypothesis [63] that the trans- influence of NO proximal ligand lability is responsible for the activation of sGC by NO [63,68,69]. The other ambivalent ligands listed in Table 6 would be expected to show analogous transinfluences and effects and this may be of particular significance in understanding the biological functions of NH2 and N2H. The electronic origins of these trans- influences are all interconnected and may be attributed to the orbital mixings similar ot those described for the nitrosyl complexes [13,61].

an

us

cr

Table 5 Trans- influences in octahedral complexes with linear and planar ambivalent ligands

ce pt

ed

M

XY(XY2) y Ru-X-Y(o) Ru-Clcis (Å) Ru-Cltrans- (Å) NO 3 180 2.469 2.381 PO 3 180 2.498 2.357 NS 3 180 2.449 2.398 SO2 2 180 2.524 2.535 N2 2 180 2.539 2.619 NH 1 180 2.398 2.349 N2H 3 172 2.384 2.479 Δ represents the calculated difference between the cis- and trans- Ru-Cl bond lengths (Å).

Δ(Å) -0.09 -0.14 -0..05 0.01 0.08 -0.05 -0.10

Ac

Table 6. Trans- influences in octahedral complexes with bent and non-planar ambivalent ligands

XY(XY2) y Pd-X-Y(o) Pd-Clcis (Å) Pd-Cltrans- (Å) NO 3 119 2.424 2.664 PH2 3 97 2.396 2.624 NS 3 119 2.419 2.599 N2H 2 118 2.419 2.581 NH2 3 96 2.405 2.540 Δ represents the calculated difference between the cis- and trans- Pd-Cl bond lengths (Å).

Δ(Å) 0.24 0.23 0.18 0.16 0.14

17

Page 17 of 34

M

an

us

cr

ip t

2.5 Ambiphilic ligands [70-87] SO2 and I2 provide specific examples of ambiphilic ligands, which display symmetry signatures. According to Green’s MLX notation [1] an uncharged ambiphilic ligand is described as L when it behaves as a Lewis base and Z when it behaves as a Lewis acid. In frontier orbital terms ambiphilic behaviour is associated with equally accessible HOMOs and LUMOs. SO2 and I2 ligands exhibit the symmetry signatures shown in Figure 10 and these alternative geometries may be interpreted either in terms of classical Lewis/VSEPR considerations, or the different symmetry properties of their donor and acceptor orbitals. The iodine molecule is not a very well studied ligand [77], but it shows linear and bent geometries, whereas SO2 exhibits planar (point group C2v) and non-planar (point group Cs) geometries [55-60,71-77].

ed

Figure 10. Alternative geometries shown by ambiphilic ligands when they co-ordinate to an acceptor or donor with a single orbital of σ symmetry.

Ac

ce pt

In the Lewis notation adducts of ambiphilic ligands are represented by a reversal of the dative bond arrow (see Figure 10). I2 acts as a Lewis base by donating from one of the lone pairs which have a predominance of p orbital character and consequently the resulting adduct has a T-shaped bent geometry. It acts as a Lewis acid by accepting donation into the I-I σ* empty orbital and thereby creating a linear three-centre two-electron bond. SO2 acts as a Lewis base by donating from its HOMO which approximates to an outpointing sp2 hybrid localised on the sulphur and as a Lewis acid by accepting an electron pair into its LUMO which is a π antibonding orbital perpendicular to the OS-O plane. This molecular orbital is localised more on S than O and consequently adduct formation results in the non-planar geometry illustrated in Figure 9 [55,56]. DFT calculations were undertaken in order to provide a deeper understanding of ambiphilic ligands. Specifically they have been used to highlight the redistribution of charge when either a main group Lewis base or an ambiphilic ligand, e.g. SO2, forms an adduct with classical Lewis acids, e.g. AlCl3 or Al(CF3)3 [2,78]. The changes in calculated Mulliken atomic charges for the adducts (summarised in Figure 11) are not entirely consistent with the classical Lewis donor bond description shown in 3 and 4. The ammonia fragment becomes more positive on adduct formation and this is balanced by the increase in charge on the SO2 and AlCl3 fragments. The stronger Lewis acidity of AlCl3 is reflected in a greater inter-fragment redistribution of charge, i.e. 0.23 vs 0.07. Interestingly the charge on the aluminium and sulphur atoms remain essentially unchanged in these adducts with NH3, and surprisingly the nitrogen atoms bears a more negative charge in the adduct than in the free molecule. This is more than compensated for by the increased positive charge on the hydrogen atoms. For AlCl3 and SO2 the retention of the same charge on the Al and S atoms is associated with a greater negative charge on Cl and O. The changes in calculated charges follow the schematic representation 18

Page 18 of 34

Ac

ce pt

ed

M

an

us

cr

ip t

shown in 2 and this does not reflect the conventional Lewis structures shown in 3 and 4, and may be better represented by the sequence of arrows shown in 5. Braunschweig [2] has noted that if attention is concentrated solely on the reversal of the dative bond direction of an ambiphilic ligand then the observable consequences are very subtle indeed and not easily detectable using spectroscopic techniques. The discussion above which emphasises the effective transfer of an electron density from donor to acceptor substituents reinforces this view. The formation of an adduct between Al(CF3)3 and SO2 shown in Figure 11 results in the transfer of 0.08e from SO2 to Al(CF3)3 and this may be compared with the 0.23e transferred from NH3 ( a better donor) to Al(CF3)3 . In summary, although SO2 behaves as an ambiphilic ligand in the examples shown in Figure 11, it is neither a strong donor or acceptor, and only 0.08 and 0.07e is donated or accepted in Al(CF3)3(SO2) and H3NSO2 respectively. For neutral ambiphilic ligands such as SO2 and I2 achieving the balance of having homos and lumos both accessible for bonding precludes their behaving as either strong Lewis bases or acids. The ambiphilic ligands SO2 and I2 also co-ordinate to transition metals and show similar symmetry signatures if the metal fragment has donor and acceptor orbitals with σ-symmetry characteristics (see Figures 12 and 13). Hoffmann and Rogachev have discussed examples of compounds where I2 functions as a Lewis base to rhodium acetate metal-metal bonded dimers, where the ligand donates to the antibonding Rh-Rh antibonding σ*-orbital[77]. Drawing on earlier work by Kubas, Moody Eller and Ryan and Mingos [55-60], van Koten [75,76] has extensively studied the Lewis acid chemistry of SO2 with platinum(II) square-planar complexes with pincer ligands. Hoffmann, and Rogachev [77] have recently published a detailed bonding analysis of the alternative bonding modes of I2 and also noted relationships with the SO2 compounds of van Koten. They have described the I2 molecule as a Janus faced ligand which displays alternative co-ordination geometries [77]. Figure 13 shows a series of related square-pyramidal SO2 and I2 adducts which have been derived from square-planar d8 complexes, which function as Lewis bases through their filled dz2 orbitals. These complexes are closely related to the bent NO+ adducts of [IrCl(CO)(PPh3)2] reported by Ibers et al [71]. The pincer ligands stabilise the square-planar Lewis base geometry relative to alternative trigonal-bipyramidal based ML4 geometries and this preference has been used to great effect by van Koten [75,76]. All the metal containing examples in Figure 13 have an EAN count of 16. When ambiphilic ligands such as SO2 and I2 form closely related or isomeric metal complexes with the alternative symmetry signatures the metal component must also be ambiphilic [8,54]. Low oxidation state transition metals of the later transition metals, e.g. [Pt(PR3)2] and [Pt(PR3)3], provide examples of such complexes (see also Table 2) and Bourissou, Braunschweig and Hill [72-74] have provided comprehensive reviews of the relevant literature. They have noted that the scope of M→Lewis acid adducts has been substantially enlarged recently by the use of “donor buttresses”, i.e. polydentate ligands which have the Lewis acid site supplemented by Lewis base donor sites. This has led to the structural characterisation of a wide range of borane, stannane, and silane M→Z complexes. In common with the nitrosyl complexes discussed above the planar and non-planar co-ordination of SO2 to transition metals is sensitive to the metal in an MLn complex. For example, the related [M(SO2)(PR3)3] and [M(SO2)(PR3)2] (M= Ni, Pd and Pt) complexes have SO2 co-ordinated with planar geometries for nickel and non-planar for palladium and platinum [79,80]. The trigonal-planar M(PR3)3 and linear M(PR3)2 molecules are ambiphilic, by virtue of an empty pz orbital which enables them to act as a Lewis acid and a filled dz2 orbital which enables it to act as a Lewis base (see Figure 14) [8,54]. The ambiphilic character of these transition metal complexes is not associated with a symmetry signature, but is nonetheless reflected in significant changes in bond angles around the metal atom. When they act primarily as a Lewis acid the geometry is that anticipated from VSEPR for a d10 ML4 or ML3 co-ordination compound, whereas when they act mainly as a Lewis base

19

Page 19 of 34

ip t cr us an M

16.

Ac

ce pt

ed

Figure 11. Calculated Mulliken charges for amine adducts of AlCl3 and SO2

Figure 12. Examples of SO2 complexes where the ligand is acting as a Lewis base. Unfortunately to date no analogous examples of mononuclear I2 complexes have been reported, but the polymeric silver complex clearly shows the ability of I2 to function as a Lewis base [81].

20

Page 20 of 34

ip t

cr

Figure 13. Examples of square-planar d8 complexes acting as Lewis bases towards SO2 and I2. [6973]

Ac

ce pt

ed

M

an

us

then the geometry closely resembles that of the ML3 and ML2 parent compound, i.e. trigonal-planar and linear. The structural data given in Figure 14 for related nickel and platinum complexes of SO2 provide specific examples of the structures which result when ambiphilic SO2 and M(PR3)x (x = 2 or 3) molecules are combined.

Figure 14. Comparison of the structures of M(PPh3)3(SO2) (M = Ni or Pt) (Bond lengths in Å) and their interpretation in terms of the frontier orbitals of the donor and acceptor molecules. Both metal orbitals will hybridise with the metal 6s orbital. Braunschweig et al have [78] recently provided an interesting comparison of SO2 and AlCl3 co-ordinated to [Pt(PCy3)2] , which defines the structural differences which result when an ambiphilic ligand and a traditional Lewis acid are co-ordinated to the same metal complex. DFT calculations on the model compounds Pt(SO2)(PH3)2 and Pt(AlCl3)(PH3)2 are summarised in Figure 15 and the geometric consequences of adduct formation are reasonably reproduced by the DFT calculations. Both structures are remarkably similar and the initially linear Pt(PCy3)2 moiety distorts away from the incoming ligand by only 14-18o and the co-ordination involves donation primarily from the dz2 orbital. 21

Page 21 of 34

ip t

cr

Figure 15. Observed (PCy3) and calculated (PH3) geometries for AlCl3 and SO2 adducts of platinum metal complexes.

Ac

ce pt

ed

M

an

us

For the SO2 adduct the bond angles and bond lengths are reasonably well reproduced although the PtS bond length is calculated to be 0.11Å longer that the experimentally observed bond length. For the corresponding AlCl3 adduct the Pt-Al bond length shows approximately the same overestimate and the P-Pt-P bond angle is considerably larger than the experimentally determined value. The bond between the platinum atom and the Lewis acids involves a filled dz2 orbital of the d10 platinum centre and consequently in its initial stage the P-Pt-S(or Al) angle is expected to be 90o [82-86]. The structure of [Ni(SO2)(PCy3)2] may be described in the new notation as {L2M(SO2) p]16 since the SO2 ligand has a planar geometry and a trigonal geometry about the nickel atom which is consistent with that expected for an ML3 d10 complex.. In contrast the corresponding palladium and platinum complexes have non planar M-SO2 geometries and a P-Pd-P angle which is much closer to 180o. The M-S bond is significantly longer in the palladium and platinum compounds and according to the new notation they would be described as {L2M(SO2) np]14. In formal valence bond terms the alternative structures may be represented the two canonical forms shown below (Figure 16), whereby the geometry at sulphur may be predicted by the VSEPR considerations and the geometry at the metal atom reflects the differences in the total electron counts.

Figure 16 Calculated bond distances and angles in model {L2M(SO2) p]16 and {L2M(SO2) np}14 compounds and estimates of the energy differences for the interconversion of the isomeric forms. The calculated energy difference between the isomeric forms of SO2 complexes, i.e.{L2M(SO2) p]16 and {L2M(SO2) np}14 (L = PH3) is particularly small for the palladium compounds, i.e. -0.6 kcal mol-1 . This energy difference may be somewhat larger for more sterically crowded phophines [85,86], and may be greater in complexes with sterically rigid spectator ligands, but the calculations on the model compounds strongly suggest a small barrier for the interconversion of the isomers. It is noteworthy that the interconversion of isomers shown in Figure 16 does not only involve a pyramidalisation at sulphur, but also a significant and simultaneous opening up of the P-MP bond angle. It follows that the relative stabilities of the isomers may be manipulated by replacing the phosphines by bidentate phosphines with variable bite angles. Calculations on model compounds 22

Page 22 of 34

71 96 125 164

2.25 2.23 2.27 2.37

Angle between M-S and SO2 plane(o) 174 176 145 113

M-SO2 geometry

cr

M-S (Å)

us

Bite angle (o)

ip t

Table 7 Effect of the ligand bite angle on the coordination mode of SO2

an

Isomeric forms of Pd(SO2)(PH3)2 P-M-P(o) M-S(Å) 117 2.22 163 2.41

planar non-planar

M

180 180 113

planar planar intermediate non-planar

Ac

ce pt

ed

with bite angles from 71 to 164o have shown that the geometry of the SO2 may indeed be manipulated by changing the bite angle . The results of these calculations are summarized in Table 7 and compared with the calculated bond angles for the isomeric forms of [Pd(SO2)(PH3)2]. Interestingly for a bite angle of 125o an intermediate geometry is observed for the M-SO2 moiety. Since the reactivity of coordinated SO2, particularly with dioxygen, has been related to its geometry [55] it follows that the abilty to manipulate the geometry of the co-ordinated SO2 ligand may also have chemical/catalytic implications [55,56] A study of the charge distribution in the isomeric palladium compounds [Pd(SO2)(PH3)2] using a Mulliken charge analysis similar to that shown in Figure 11 for main group adducts suggests that the two isomers result in almost the same transfer of electron density from the metal phosphine complex to SO2, suggesting that the simple Lewis description in terms of a reversal of the dative bond is an oversimplification. The synergic bonding interactions between orbitals on SO2 and the M(PH3)2 fragments appear to equalize the transfer of charge and this leads to a very similar distribution of electron density in the isomeric forms of the adducts. The calculated enthalpies for adduct formation with [Pt(PH3)2] for the ambiphilic ligands are compared with those of other ligands in Figure 17. The AlCl3 adduct [78,87] is significantly more stable than the SO2 adduct and the enthalpy change is comparable to that of PH3. The most stable adduct is formed by CO. The Figure also gives the calculated enthalpies changes for [H 3NSO2] and [H3NAlCl3] – for the former the enthalpy change is less favourable than the platinum analogue, whereas for the latter the enthalpy change is more favourable. The limited data available indicates that the hard and soft acid-base theory is transferable to ambiphilic and Lewis acidic ligands coordinated to platinum complexes which are soft Lewis bases. The corresponding calculations on the related palladium complexes show that the adduct formation of Pt(PH3)2 with Lewis acids and bases follow the same order as noted above, but the absolute values are smaller, suggesting that the platinum complex is a better Lewis base. 23

Page 23 of 34

ip t cr us

an

Figure 17. Calculated enthalpies of adduct formation for Lewis acid and Lewis base adducts of [Pt(PH3)2] and NH3.

Ac

ce pt

ed

M

Co-ordination of these Lewis acid ligands perturbs the metal-ligand geometry to a lesser extent than a classical Lewis base, which would be expected to lead to bond angles closer to 120 o characteristic of a d10 ML3 complex. It is noteworthy that the ambiphilic [M(PR3)3] and [M(PR3)2] (M= Ni and Pt) molecules do not display symmetry signatures for ambivalent and ambiphilic ligands such as SO2 and NO, but nonetheless the differences in the P-M-P angles do provide an angular signature. The bond lengths between [M(PR3)3] and [M(PR3)2] and SO2 also provide circumstantial evidence for the alternative bonding modes, with the non-planar geometries generally showing longer M-S bond lengths [55,56]. A comparison of SO2 (6 and 7) and related NO complexes suggests [55,56] that the barrier for converting SO2 from planar to non-planar is smaller than that for converting linear to bent NO. This conclusion has been supported by calculations by Kubas, Moody and Ryan [55-57,59,60] who have related the more facile conversion of planar to non-planar SO2 to the smaller HOMO-LUMO gap in SO2 compared to NO [79-80] . The SO2 ligand is not only ambiphilic but also ambidentate and may also act a two electron donor by co-ordinating to a transition metal in a π-bonded fashion and the subtle interplay between σ and π forms of the ligand have also been studied structurally and been the subject of theoretical studies [84-87]. The ambivalent character of the [M(PR3)2] molecule is retained when two SO2 ligands are coordinated to them as indicated in the structures illustrated below. The larger P-Pt-P bond angle and the non pyramidal symmetry equivalent SO2 ligands in 7 suggest that the platinum complex behaves as a Lewis base whereas the nickel complex acts as a classical Werner Lewis acid. It is noteworthy that in the second example a planar/non-planar structure is not adopted in the platinum complex.

24

Page 24 of 34

Ac

ce pt

ed

M

an

us

cr

ip t

Summary This review has drawn heavily on previously articulated ideas that transition metals and main group unsaturated molecules may function as either Lewis acids or Lewis Bases, but breaks new ground in proposing that there is an important sub-classes of ambivalent and ambiphilic molecules which include transition metals and main group molecules (or ions) which adopt alternative geometries. It is proposed that these be described as ambivalent or ambiphilic Lewis acids/bases with symmetry signatures. The symmetry signatures may be observed only at the ligand, at the metal, or at both the ligand and the metal. By drawing attention to the common characteristics of these ligands it is hoped that further experimental work may be encouraged. In the absence of such experimental data, DFT calculations have been used to provide an initial insight into the relative properties of ambivalent and ambiphilic ligands. Specifically, the following points of interest have been highlighted by the theoretical studies. 1. The relative abilities of ambivalent and ambiphilic ligands to adopt the less symmetric geometry follows the order PO > PH2 > N2H > SO2 > NO > NH2 > NS in square-pyramidal complexes. 2. The isomeric forms of the complexes with linear and bent geometries (or planar and nonplanar) have very similar energies and calculations on model compounds suggest the energy difference is small, i.e. approximately 5 kcalmol-1. 3. Complexes containing both linear and bent (or planar and non-planar) ambivalent ligands may undergo a low energy valence tautomerism based on the interconversion of the two forms. Such complexes may be in equilibrium with an isomer with both ligands linear (or planar). 4. The more symmetrical forms of the ambivalent ligands are not associated with large transinfluences – indeed some show negative trans- influences, The less symmetrical forms of the ambivalent ligands show a high trans- influence and the relative ordering is NO > PH2 > NS > N2H > NH2 . The high trans- influence in these complexes has chemical and biological implications. Furthermore, the presence of a ligand with a high trans- influence may also lead to the isolation of pairs of octahedral and square- pyramidal complexes with 18 and 16 electrons respectively. Many examples of such compounds have been noted for nitrosyl and nitrido- complexes. 5. The study of transition metal complexes with ambiphilic ligands, e.g. SO2, has underscored the view that the resultant adducts are associated with ambiphilic metal complexes and isomeric forms arise from changes in geometry at the ligand and the metal. 6. The calculations suggest that the relative stabilities of platinum complexes with ambiphilic and related ligands follows the order of stability CO > AlCl3 > PH3 > SO2 > NH3. 7. The geometries of ambiphilic ligands with symmetry signatures may be manipulated by changing the bite angles of the ligands also co-ordinated to the metal. The isomeric forms of adducts of ambiphilic ligands are associated with large changes in the bond angles around the

25

Page 25 of 34

metal and constraining these angles using bidentate ligands preferentially stabilises one of the isomers. The identification of the essential properties of these classes of ligands is important for understanding the structural and catalytic properties of metal complexes based on these ligands and may have relevance to the ability of some of these molecules to function as gasotransmitters.

ip t

Appendix Notation for ambivalent and ambiphilic ligands In 2013 [7,8] I introduced a new notation for ambivalent and ambiphilic ligands and the

ed

M

an

us

cr

essential features are summarised below. The vast majority of complexes of ambivalent ligands conform to the 18 and 16 electron rules and the new notation focuses attention on the total electron count rather than the modified d electron count proposed by Enemark and Feltham for transition metal nitrosyls [88]. The routine determinations of single crystal X-ray structures of co-ordination compounds these days also means that a structural designator can be incorporated into the notation using the abbreviated symbols introduced in Table 1, i.e. for diatomic ligands (XY), which are capable of taking up M-X-Y bond angles between 180 and 110o: 180-160o l (linear); 140-160o i (intermediate); 110-140o b (bent). For XY2 ligands the corresponding structural designators are : planar (p), intermediate (i) and non-planar (np). For these ligands this classification may be based on the dihedral angle between the Y-X-Y plane and the M-X bond or the sum of the bond angles at the donor atom, X, ΣDα :360-345o for p, 345-335o for i and 335-320o for np [ ]. The proposed new notation takes the following forms for some common complexes:

Ac

ce pt

Linear single bent and planar [LmM(NO)n l]y [LmM(N2R)n sb]y [LmM(SO2)n p]y Intermediate [LmM(NO)ni]y [LmM(N2R)n i]y [LmM(SO2)n i]y Bent, double bent and non-planar [LmM(NO)n b]y [LmM(N2R)n db]y [LmM(SO2)n np]y where n + m = coordination number of complex and y = EAN count with the ambivalent ligand donating the number of electrons defined in Table 1 for the alternative geometries (e.g. NO donating 3 electrons for l , and 1 electron for b and either 3, or 1 for i (intermediate)). For complexes with intermediate geometries, e.g. M-N-O angles between 140 and 160o, the designated number of electrons donated is determined by supplementary spectroscopic, magnetic and structural data. Chemical precedence based on the 18 and 16 electron rules will also assist the assignment of the appropriate classification. Figure 2 provides some illustrative examples of this notation, For odd electron donors such as NO, N2R etc the well documented semantic complications which result from formally designating them as either L+ and L- is circumvented by formally and consistently using the electron donor characteristics of the ligand in its neutral from, i.e. the electron counts summarised in Tables 1 and 2. The formal oxidation state of the metal is not assigned and the molecules are classified according to the total number of electrons donated to the central metal atom. Numerous DFT calculations on complexes of the ligands given in Table 1 [13-33] have underlined the validity of the electroneutrality principle and therefore this seems the most reasonable approximate starting 26

Page 26 of 34

M

an

us

cr

ip t

point for describing the bonding in classical Lewis electron pair bonding representations. In N2R complexes the R group does not lie along the MNN rotation axis and therefore the alternative geometries are commonly described as bent(b) and double-bent(db). The N2R ligand is also capable of bonding in a π-fashion (η2) (see Figure 1) and therefore is both ambidentate and ambivalent. The Table also provides examples of ambivalent ligands which donate either 4 or 2 electrons. The alkoxy ligand, OR, is particularly flexible and is capable of donating 5, 3 or 1 electrons. The series of tetrahedral compounds OsOn(NBut)4-n , which have been structurally characterised, may be classified as follows in the new notation: {L3M(NBut) l}18 , {L2M(NBut)2 lb}18 , {LM(NBut)3 lbb}20 , {M(NBut)4 iiii}22 . It is not uncommon for high symmetry oxo-, nitrido- and imido complexes to exceed the EAN rule particularly when they have high symmetries. The additional electrons occupy non-bonding orbitals which are localized on the ligands. This aspect has been discussed extensively by Hall, Schrock, Kaltsoyannis and Mountford [17,21,32]. In {LM(NBut)3 lbb}20 {L2M(NBut)2 lb}18 the presence of linear and bent forms of the alkylimido- simultaneously have been rationalized in terms of the EAN Rule [37] and give rise to the possibility of valence tautomerism. There are also examples of alkylimido- complexes where rather than showing a linear-bent duality both ligands adopt intermediate geometries [28]. In view of the many disputes which have arisen in the literature as a result of the assignment of a formal charge to NO and the formal oxidation state or valency of the metal, which follows [88], a notation which is not dogmatic and which specifies only the total number of valence electrons has many advantages[7,8].

Ac

ce pt

ed

Computational Details All DFT calculations were carried out using the Amsterdam Density Functional package, ADF2013, and the results refer to the gas state and solvent effects were not taken into account [8995]. The numerical integration scheme applied for the calculations was developed by te Velde et al.; the geometry optimization procedure derives from that of Versluis and Ziegler. Geometry optimizations were carried out using the local density approximation of Vosko, Wilk, and Nusair (LDA VWN) augmented with the nonlocal gradient correction from Perdew and Wang. Relativistic corrections were added using a scalar-relativistic zeroth order relativistic approximation (ZORA) Hamiltonian for all the transition metal atoms after initial trial calculations on related Ni, Pd and Pt complexes. The electronic configurations of the molecular systems were described either by a triple-ζ with single polarization (TZP) basis set for all metal atoms or by a mixed basis set where the TZP basis set was used on the metals, S, O, Al and P, while a double- ζ basis set (DZ) was used on C and H. Non-hydrogen atoms were assigned a relativistic frozen core potential, treating as core the shells up to and including the following: 1s {O, C, N}, 2p {first-row transition metals, P, S}, 3d {secondrow metals}, and 4d {third-row metals}. A set of auxiliary s, p, d, and f functions, centered on all nuclei, was used to fit the molecular density and represent Coulomb and exchange potentials accurately in each SCF cycle. The Figures below compare the results of these calculations for [Ru(NO)Cl5]2- (GGA.OPBE exchange correlation potential with scalar relativity corrections and the basis set choices noted above) with the solid state X-ray crystallographic data of Veal and Hodgson and Emel’yanov et al [96,97,98]. The bond lengths associated with the Cltrans-Ru-N-O moiety is well reproduced and the Ru-Clcis- bond lengths are longer than the experimental values by 0.08Å, but the inverse trans- influence observed for the linear NO ligand is reproduced. Although the theoretical results on the [Ru(NO)Cl5]2- anion refers to the gas phase they reasonably reproduce the bond length 27

Page 27 of 34

cr

ip t

variations in the solid state shown below for [NH4]2[Ru(NO)Cl5] and K2[Ru(NO)Cl5]. Since the calculations were used to study and highlight the relative trans- influences of ambivalent ligands with their alternative symmetry signatures this level of approximation was deemed to be satisfactory.

Ac

ce pt

ed

M

an

us

The corresponding calculations for the complexes with the bent and non-planar geometries which have a total of two additional valence electrons were completed on [Pd(XY)Cl5]2- and [Pd(XY2)Cl5]2in order to maintain the same charge on the complex anion. Since the study involved the development of some general trends associated with transition metal complexes of ambivalent and ambiphilic ligands the bulky aryl and alkyl phosphine ligands were modeled by PH3. In the context of SO2 complexes Gilbert [85] has analysed in some detail the the extent to which this approximation is justified and also demonstrated how the electronic and steric consequences of the more bulky ligands may be more satisfactorily modeled using a combination of DFT and molecular mechanics methodologies. Mulliken overlap populations have their limitations and references provide some alternative procedures which have been used for analyzing the bonding in metal nitrosyl complexes. In the present context they have been used to give the relative donoracceptor properties of ambivalent and ambiphilic ligands rather than the precise charges on specific atoms and therefore largely for pedagogical reasons they have been used for interpreting the results for an audience of synthetic chemists. The Table below summarises the results of different basis set on the calculated geometries of a the model compound [RuCl(NO)2(PH3)2]+ The question of using the appropriate basis sets for analysing the bending of the nitrosyl ligand in transition metal nitrosyl complexes has been discussed by several groups [18,19,26,38,41,99,100] have demonstrated that the calculated bond lengths and angles were not very sensitive to the choice of basis set. The Table below confirms that for our model system the OPBE option was used for the studies described in the review above. All the calculations summarised below reproduced the following features of the structure of [RuCl(NO) 2(PPh3)2]+ [38,41]: the square pyramidal geometry at the ruthenium atom, a bent nitrosyl geometry for the apical nitrosyl ligand with Ru-N-O approximately 120o, a linear nitrosyl geometry for the basal nitrosyl with Ru-N-O approximately equal to 180o; the bent nitrosyl eclipsing the linear Ru-N-O bond. The BP86-d option was not used because it calculates the alternative bent isomer with NO bent towards chorine as more stable. Although the bulky phosphines ligands were modelled by PH3 the RuP bond lengths are reproduced pretty well in agreement with previous analyses on SO2 complexes of M(PHxPh3-x) (M = Ni,Pd and Pt) by Gilbert [85].

28

Page 28 of 34

OPBE 1.929 1.78 1.16 1.154 2.398 2.404

LDA 1.897 1.784 1.157 1.151 2.387 2.386

PBE 1.954 1.807 1.168 1.162 2.438 2.443

OLYP 1.965 1.799 1.165 1.16 2.453 2.455

BLYP 1.989 1.825 1.173 1.168 2.491 2.496

ip t

X-ray 1.853 1.743 1.166 1.158 2.431 2.419

cr

Ru-N1(axial) Ru-N2(equa) N1-O1 N2-O2 Ru-P1 RuP2

Ac

ce pt

ed

M

an

us

The primary aim of the present analysis is to demonstrate that there exists a broad range of ligands which can behave in an ambivalent or ambiphilic manner and they exhibit symmetry signatures. Therefore the angular characteristics of the ligand are the primary focus of the study and the frequent comparisons between structural data and the theoretical analyses presented suggest that the calculations appropriately reflect the variations reasonably accurately. Future, studies will no doubt make more detailed studies of specific examples which will evaluate in more detail and account more satisfactorily for the subtle effects which arise from the fact that the phosphine ligands are not pure spectator ligands. Specifically future work on the very soft potential energy surfaces for the tautomeric and isomeric transformations of the complexes will need to focus greater attention on the electronic and steric effects associated with the alkyl and aryl phosphine ligands. The enthalpy differences for adduct formation were estimated by subtracting the sum of the energies of the LnM fragment and of SO2, NH3 or AlCl3 from the energy of the LnM(SO2) molecule. The geometries of the adduct and the parent molecules were optimized with the OPBE functional . The data were not corrected for basis set superposition error (BSSE), because the correction at this basis set level is probably 2.0 kcal mol-1 and because it is probably systematic across the series of molecules investigated, and thus will not affect comparisons. The bidentate ligands were modeled by (H2P)2CnH2n (n = 1, 3 5) and (H2P)2C10H4, where the large bite angle was generated by the alkyne spacers shown below:

Acknowledgements I should like to thank Professors Hoffmann, Parkin, Mountford and McGrady for many helpful discussions and their incisive comments.

29

Page 29 of 34

M

an

us

cr

ip t

References [1] G. Parkin, Comprehensive Organometallic Chemistry III, ed R.H. Crabtree and D.M.P. Mingos, (2006) vol.1 , ch 1, Elsevier, Oxford, UK; G. Parkin, J. Chem. Educ. 83 (2006) 791; M.L.H. Green, J. Organometal. Chem. 500 (1995) 127. [2] J. Bauer, H. Braunschweig and R.D. Dewhurst, Chem. Rev. 112 (2012) 4329. [3] J.E. Huheey, E.H. Keiter and R.L. Keiter, “Inorganic Chemistry - Principles, Structure and Reactivity”, Harper Collins College Publishers, New York, USA (1993) 197. [4] F.A. Cotton, G. Wilkinson and P.L. Gaus, “Basic Inorganic Chemisry”, 3rd Edition, John Wiley and Sons, Chichester, UK, (1995) 828. [5] P. Atkins, T. Overton, J. Rourke, M. Weller and F.A. Armstrong, “Shriver and Atkins Inorganic Chemistry”, 5th Edition, Oxford Unversity Press, Oxford, UK, (2010), 410. [6] L. Pauling “Nature of the Chemical Bond”, Cornel University Press, 2nd. Ed, Cornell, New York, USA (1947) 251. [7] D.M.P. Mingos, J. Organomet. Chem. 751 (2013) 153. [8] D.M.P. Mingos, “Nitrosyl Complexes in Chemistry, Biology and Medicine”, Struct. Bond. 154 (2014),1; 155 (2014) 1. [9] A. Butler and R. Nicholson (2003) “Life Death and Nitric Oxide”, Royal Society of Chemistry, Cambridge, UK [10] R. Wang (2004) “Signal Transduction and the Gasotransmitters NO, CO and H2S in Biology and Medicine”, Humana Press, Totowa, New Jersey, USA; P.K. Allan, R. Morris, Struct. Bond.154 (2014) 225. [11] E.H. Thorsteinson, F. Basolo, J. Amer. Chem. Soc. 88 (1966) 3929, D.P. Keeton, F.

Basolo, Inorg. Chim. Acta, 6 (1972) 33. [12] Y. Jiang and H. Berke, Struct. Bond., 153 (2014) 167.

Ac

ce pt

ed

[13] R. Hoffmann, M. Elian, M.M.L. Chen, A.R. Rossi and D.M.P. Mingos, Inorg. Chem. 13 (1974) 2666. [14] S. Kahlal, J-Y. Saillard, J-R. Hamon, C. Manzur and D. Corillo, New. J. Chem. 25 (2001) 231. [15] M.T. Benson, J.C. Bryan, A.K. Burrell and T.R. Cundari, Inorg. Chem. 34 (1995) 2348; M.T. Benson, T.R. Cundari and E.W. Moody, (1995) J. Organometal. Chem. 504 (1995)1. [16] C. Lyubinova, O.V. Sizova, C. Losden and G. Frenking, J. Molecular Struct. THEOCHEM 865 (2008) 28. [17] N. Kaltsoyannis and P. Mountford, J. Chem. Soc. Dalton Trans. (1999) 781; A.D. Schwarz A.J. Nielson, N. Kaltsoyanis and P. Mountford, Chemical Sciences 3 (2012) 819 [18] T.C. Berto, V.C.C. Praneeth, I.C. Goodrich and N. Lenhert, J. Amer. Chem. Soc. 131 (2009) 17116. [19] N. Lenhert, J.T. Sage, N.J. Silvernail, W.R. Scheidt, E.E. Alp, W. Sturhahn and J. Zhao, Inorg. Chem. 49 (2010) 7197; G.B. Richter-Addo, R.A. Wheeler, C.A. Hixson, L. Chen, M.A. Khan, M.K. Ellison, C.E. Schulz and W.R. Scheidt, J. Amer. Chem. Soc. 123 (2001) 6314. [20] R. Hoffmann and D.L. DuBois, Nouveau J. Chem. 1 (1977) 479. [21] Z. Lin and M.B. Hall, Coord. Chem. Rev. 123 (1993)149. [22] M.R. Russo, N. Kaltsoyannis and A. Sella, Chem. Commun. (2002) 2458 [23] T.A. Manz, A.E. Fenwick, H. Phomberi, I.P. Rothwell and K.T.Thompson J. Chem. Soc. Dalton Trans. (2005) 668 [24] M.T. Benson, T.R. Cundari and E.W. Moody, J. Organometal. Chem. 504 (1995)1. [25] P.D. Lyne and D.M.P. Mingos, J. Chem. Soc. Dalton Trans. (1995) 1635. [26] A. Ghosh, Biolog Inorg. Chem. 11 (2006) 671; J. Conradie, A. Hosh and A. Ghosh, J. Phys. Chem. B 111 (2007) 12621 30

Page 30 of 34

Ac

ce pt

ed

M

an

us

cr

ip t

[27] D.M.P. Mingos, P.D. Lyne and U. Pidun, Pure Appl. Chem. 67 (1994) 2651. [28] D.J. Rankin, H.E. Robertson, A.A Danapoulos, P.D. Lyne, D.M.P. Mingos and G. Wilkinson, J. Chem. Soc. Dalton Trans. (1994) 1563. [29] M.T. Benson, J.C. Bryan, N.C. Burrell and T.R. Cundari, Inorg. Chem. 34 (1995) 2348. [30] J. Conradie, K.H. Hopmann and A. Ghosh, J. Phys. Chem. B 114 (2010) 8517. [31] P. Hummel and H.B. Gray, Coord. Chem. Rev. 203 (2000) 5. [32] M.H. Schofield, T.P. Kee, J.T. Anhaus, R.R. Schrock, K.H. Johnson and W.M. Daws, Inorg. Chem. 30 (1991) 3595. [33] K. Tatsumi and R.Hoffmann, Inorg. Chem. 21(1982) 466; B.C. Ward and J.L. Templeton, J. Amer. Chem. Soc, 102 (1980) 3288,. [34] D.C. Brower, D.M.P. Mingos and J.L. Templeton, J. Amer. Chem. Soc. 109 (1987) 5203. [35] W.A. Nugent and J.M. Mayer, “Metal-Ligand Multiple Bonds”, Wiley Interscience, New York, USA (1988). [36] N Kaltsoyannis and J.E. McGrady, Principles and Applications of Density Functional Theory in Inorganic Chemistry I, Structure and Bonding, 112 (2004) 1; Principles and Applications of Density Functional Theory in Inorganic Chemistry II, Structure and Bonding, 113 (2004) 1. [37] M.V. Putz and D.M.P. Mingos, Applications of Density Functional Theory to Chemical Reactivity, Structure and Bonding, 149 (2012) 1; Applications of Density Functional Theory to Biological and Bioinorganic Chemistry, Structure and Bonding, 149 (2012) 1. [38] C.G. Pierpont and R. Eisenberg (1971) J. Amer. Chem. Soc. 93 (1971) 4905. [39] R.T. Baker, J.F. Whitney and S.S. Wreford, Organometallics, 2 (1983) 1049 [40] R.T. Baker, P.J. Krusic, T.H. Tulip, J.C. Calabrese, S.S. Wreford, J. Amer. Chem. Soc., 105 (1983) 6763. [41] C.G. Pierpont and R. Eisenberg, Inorg. Chem. 12 (1972) 1088. . [42] D.J. Sherman, D.M.P. Mingos and I.D. Williams, Trans. Met. Chem., 12 (1987) 405. [43] J.H. Waters and K.R. Whittle, J. Chem. Soc. Chem. Commun. (1971) 518. [44] M.W. Schonerer, E.C. Baker and R. Eisenberg, J. Amer. Chem. Soc. 103 (1979) [45] J.P. Collman, P.Farham and G. Dolcelli, J. Amer. Chem. Soc. 93 (1971) 1788. [46] L.K. Bell, J. Mason, D.M.P. Mingos and D.G. Tew, Inorg. Chem., 22 (1983) 3497. [47] D.H. Evans, J. Mason, D.M.P. Mingos and A. Richards, J. Organometal. Chem., 249 (1983) 293. [48] J. Mason, D.M.P. Mingos, D.J. Sherman, J. Schaffer and E.O. Stejskal, J. C. S. Chem. Commun. (1985) 444. [49] J. Mason, D.M.P. Mingos, D.J. Sherman and R.W.M. Wardle, J. Chem. Soc. Chem. Commun., (1984) 1223. [50] D.M.P. Mingos, D.J. Sherman, Adv. Inorg. Radiochem. 34 (1989) 293. [51] S. Alvarez and M. Liunell, J. C. S. Dalton Trans. (2000) 3288 [52] D.C. Bradley, S.R. Hodge, J.D. Runnades, M. Hughes, J. Mason and R.L. Richards, J. Chem. Soc. Dalton Trans., (1997) 1603. [53] B.L. Haymore, M. Hughes, J. Mason and R.L. Richards, J. C. S. Chem., Commun., (1988) 2395. [54] D.M.P. Mingos, J. Organometal. Chem., 689 (2004) 4420; 691 (2006 3165. [55] R.R. Ryan, G.T. Kubas, D.C. Mood and P.G. Eller, Struct. Bond. 46 (1981) 47. [56] D.M.P. Mingos, Trans. Met. Chem. 3 (1978) 1. [57] D.C. Moody and R.R. Ryan, Inorg. Chem. 15 (1976) 1823. [58] D.M.P. Mingos and D.J. Sherman, Trans. Met. Chem., 12 (1987) 400. [59] P.G. Eller, D.C. Moody and R.R. Ryan, Inorg. Chem. 15 (1976) 2442.

31

Page 31 of 34

Ac

ce pt

ed

M

an

us

cr

ip t

[60] P.G. Eller, D.C. Moody and R.R. Ryan, Inorg. Chem. 18 (1979) 223; R.R. Ryan and P.G. Eller, Inorg. Chem. 15 (1976) 495. [61] D.M.P. Mingos, Inorg. Chem., 10 (1974) 1479. [62] M.L. Tobe, “Inorganic Reaction Mechanisms”, Thomas Nelson and Sons, London, 1972 [63] T.G. Traylor and V.S. Sharma, Biochemistry 31 (1992) 2847 [64] M.F. Perutz, J.V. Kilmartin, K. Nagai, A. Szabo, and S.R. Simon, Biochemistyr, 15 (1976) 378. [65] G.R.A. Wyllie and W.R. Scheidt, Chem. Rev., 102 (2002) 1067. [66] D.M. Copeland, A.S. Soares, A.H. West and G.B. Richter-Addo, J. Inorg. Biochem., 100 (2006)1413. [67] W.R. Scheidt, K. Hatano, G.A. Rupprecht and P.L. Piciulo, Inorg. Chem., (18 (1979) 292. [68] E.A. Dierks, S. Hu, K.M. Vogel, A.E. Yu, T.G. Spiro and J.N. Burstyn, 119 (1197) 7316. [69] E.R. Derbyshire and M.A. Marletta, Ann. Rev. Biochem., 81 (2012) 559. [70] L. Vancea, M.J. Bennett, C.E. Jones, R.A. Smith and W.A.G. Graham, Inorg. Chem. 16 (1977) 97. [71] J.A. Ibers and K.W. Muir, Inorg. Chem. 8 (1969) 1921. [72] A. Amgoune and D. Bourissou, J. C. S. Chem. Commun. 47 (2011) 859. [73] H. Braunschweig H, K. Grass and K. Radacki (2005) Angew Chem Int Ed 46:7782. [74] I.R.A. Crossley, A.F. Hill and A.C Willis, Organometallics 25 (2006) 289. [75] G. Van Koten, Pure Appl. Chem. 62 (1990) 1155; M. Albrecht, M. Lutz, A.L. Spek and G. van Koten, Nature 406 (2000) 970. [76] S. Bonnet, J.H. van Lenthe, H.J.J. van Dam, J.J. Hubertus, G. van Koten, G. Klein and J.M. Robertus, J. C. S. Dalton Trans. 40 (2011) 2542. [77] A Yu Rogachev and R. Hoffmann, J. Amer. Chem. Soc. 135 (2013) 3762. [78] H. Braunschweig, K. Gruss and K. Radaccki, Angew. Chem. Int. Ed. 46 (2007) 7782; Inorg. Chem 47 (2008) 8595. [79] J.M. Ritchey, D.C. Moody and R.R. Ryan, Inorg. Chem. 22 (1983) 2276. [80] S. Arifhodzic-Radojevic, A.D. Burrows, N. Choi, M. McPartlin, D.M.P. Mingos, S.V. Tarlton and R. Vilar, J Chem. Soc. Dalton Trans. (1999) 3981 [81] T.S. Cameron, J. Passmore and X. Wang, Angew. Chem. Intern. Ed. 43 (2004) 1995. [82] I.P. Romm, Yu. G. Noskov and A.A. Malkov, Russ. Chem. Bull. Internat. Ed., 56 (2007) 1935 [83] J-G. Shim, Y.H. Jhon, J-H. Kim, K-R. Jang and J. Kim, Bull. Korean Chem. Soc., 28 (2007) 1609 [84] J.J. Oh, H.K. Hillig and R.L. Kuczkowski, J. Amer. Chem. Soc. 113 (1991) 7480. [85] J. Petzer and T.M. Gilbert, Inorg. Chem. 42(2003) 7207;D.J. Brust and T.M. Gilbert, Inorg. Chem. 43 (2004)1116.; [86] S. Sasuku, H. Sato, Y. Imai and Y. Morokuma, Inorg. Chem. 24 (1985) 4538. [87] C. Goedecke, P. Hillebrecht, T. Uhlemann, P. Haunschild and G. Frenking, Canad. J. Chem. 87 (2009) 1470. [88] J.H. Enemark and R.D. Feltham, Coord. Chem. Rev. 13 (1974) 339. [89] The ADF programmes are discussed in D. Young Computational Chemistry, WileyInterscience, 2001. Appendix A. A.2.1 332. The programmes are available from Scientific Computing and Modeling , 2013-01d released January 2013 www.scm.com. ADF2013, SCM, Theoretical Chemistry, Vrije Universiteit, Amsterdam, The Netherlands, http://www.scm.com. [90] G. te Velde, F. M. Bickelhaupt, S. J. A. van Gisbergen, C. Fonseca Guerra, E. J. Baerends, J. G. Snijders and T. Ziegler, J. Comput. Chem., 2001, 22, 931. 32

Page 32 of 34

ip t

cr

us an M

[99] [100]

ed

[96] [97] [98]

ce pt

[92] [93] [94] [95]

C. Fonseca Guerra, J. G. Snijders, G. te Velde and E. J. Baerends, Theor. Chem. Acc., 1998, 99, 391; G. De Velde and E.J. Baerends, J. Comput. Chem., 99 (1992) 84. I. Versluis and T. Ziegler, J. Chem. Phys. 88 (1988) 322. J.G. Snijders, E.J. Baerends and P. Ros, Mol. Phys., 38 (1979) 1909. E. van Lethe, R. Van Leeuwenm, E.J. Baerends and J.G. Snijders, Int. J. Quantum.. Chem., 57 (1996) 281. J.T. Veal and D.J. Hodgson Inorg. Chem., 11 (1972) 1420. J.T. Veal and D.J. Hodgson, Acta Cryst, B28 (1972) 3525. V.A. Emel’yanov, S.A. Gromilov and I.A. Baidina, A.V. Virovits, A.V. Beryaev and V.A. Logvinenko, J. Struct. Chem., 40 (200) 883. V.I. Baronovski, O.V. Sizova, J. Struct. Chem., 49 (2008) 803 O.V. Sizova, A. Yu. Sokolov, L. V. Skripnikov, O.O. Lyubimova, J. Struct. Chem., 48 (2007) 28

Ac

[91]

33

Page 33 of 34

Mingos Ambivalent and ambiphilic ……Highlights

ip t

cr

us



an



M



ed



This review defines clearly ambidentate, ambivalent and ambiphilic ligands DFT calculations have been completed which define the relative abilities of ligands to adopt non-linear and non-planar geometries. The trans- influences of linear and planar and bent and non-planar ambivalent and ambiphilic ligands are compared using DFT calculations The energetics of valence tautomerism in nitrosyl SO2 and phophido- complexes are discussed. The energetics of SO2 , AlCl3 and NH3 and CO adducts of zero oxidation state complexes of platinum phosphine complexes have been calculated and discussed. The role of the bite angle of the spectator ligand on the relative stabilities of the isomeric forms of SO2complexes has been evaluated.

ce pt



Ac



Page 34 of 34